Estimation of erosion in tectonically active orogenies. Example from the Bhotekoshi catchment, Himalaya (Nepal)

Abstract

The inaccessibility of active orogens such as the Himalayas renders detailed field studies on a large scale impossible. The integrative application of several remote sensing techniques in combination with ground truth information provides an alternative to assess the acting processes during erosion in tectonically active orogens. This study was carried out in the catchment area of the Bhotekoshi River, which drains both the Tibetan Plateau and the Himalayan Range. On the basis of stereogrammetric digital elevation models (DEMs) extracted from ASTER data and from aerial photographs, in combination with multispectral analysis, it is possible to appraise the nature of erosive processes. River networks are extracted from DEMs and their gradient, channel densities and flow path geometries are calculated and integrated in the erosion estimation model. In this study a modified Revised Universal Soil Loss Equation (RUSLE) and the Vegetation cover–Surface material and Drainage density (VSD) index have been tested for their suitability in the Himalayas. Using classifications methods – Maximum Likelihood (MLC), Support Vector Machines (SVM) and Object Based (OBC) – precise land use mapping can be carried out and affiliated with typical erosion rates. Long‐term total erosion has been estimated using power law stream analysis. We show that the erosion models tested here underestimate erosion in tectonically active regions.

Acknowledgments

We thank all the persons and institutions which contributed to the realization of this work. First we would like to thank the school partnership program NAMASTE‐S GmbH for their logistical support. Thanks go also to Sanam Aksha (Raj) for being such a good field guide, and the College of Applied Sciences, Nepal for the sharing of important information. We would like to thank Isabelle Providoli from ICIMOD for support and the numerous contacts. Also special thanks go to Roshan Bajracharya from the Kathmandu University for kindly providing us with a current flow meter and Mr. Sthapit from the Bhote Koshi Power Company PVT. Ltd for sharing data. Last but not least we have to thank the Remote Sensing Group at the Technical University of Freiberg for the supportive collaboration. Especially we have to mention Claudia Hahn for her support with SVM, Prashanth Marpu for the help with OBC and Matlab, Faisal Shahzad for his help with Matlab, Maria Helbig with map projection, Florian Wobbe for everything concerning computers, Mathias Leidig for his work on river profiles and Michael Buchwitz for supportive corrections. Mr. Saraf is thanked for his constructive review.